The application of directional microphones to hearing aids is well known in the patent literature (Wittkowski, U.S. Pat. No. 3,662,124 dated 1972; Knowles and Carlson, U.S. Pat. No. 3,770,911 dated 1973; Killion, U.S. Pat. No. 3,835,263 dated 1974; Ribic, U.S. Pat. No. 5,214,709, and Killion et al. U.S. Pat. No. 5,524,056, 1996) as well as commercial practice (Maico hearing aid model MC033, Qualitone hearing aid model TKSAD, Phonak "AudioZoom" hearing aid, and others).
Directional microphones are used in hearing aids to make it possible for those with impaired hearing to carry on a normal conversation at social gatherings and in other noisy environments. As hearing loss progresses, individuals require greater and greater signal-to-noise ratios in order to understand speech. Extensive digital signal processing research has resulted in the universal finding that nothing can be done with signal processing alone to improve the intelligibility of a signal in noise, certainly in the common case where the signal is one person talking and the noise is other people talking. There is at present no practical way to communicate to the digital processor that the listener now wishes to turn his attention from one talker to another, thereby reversing the roles of signal and noise sources.
It is important to recognize that substantial advances have been made in the last decade in the hearing aid art to help those with hearing loss hear better in noise. Available research indicates, however, that the advances amounted to eliminating defects in the hearing aid processing, defects such as distortion, limited bandwidth, peaks in the frequency response, and improper automatic gain control or AGC action. Research conducted in the 1970's, before these defects were corrected, indicated that the wearer of hearing aids typically experienced an additional deficit of 5 to 10 dB above the unaided condition in the signal-to-noise ratio ("S/N") required to understand speech. Normal hearing individuals wearing those same hearing aids might also experience a 5 to 10 dB deficit in the S/N required to carry on a conversation, indicating that it was indeed the hearing aids that were at fault. These problems were discussed by applicant in a recent paper "Why some hearing aids don't work well!!!" (Hearing Review, Jan. 1994, pp. 40-42).
Recent data obtained by applicant and his colleagues confirm that hearing impaired individuals need an increased signal-to-noise ratio even when no defects in the hearing aid processing exist. As measured on one popular speech-in-noise test, the SIN test, those with mild loss typically need some 2 to 3 dB greater S/N than those with normal hearing; those with moderate loss typically need 5 to 7 dB greater S/N; those with severe loss typically need 9 to 12 dB greater S/N. These figures were obtained under conditions corresponding to defect free hearing aids.
As described below, a headworn first-order directional microphone can provide at least a 3 to 4 dB improvement in signal-to-noise ratio compared to the open ear, and substantially more in special cases. This degree of improvement will bring those with mild hearing loss back to normal hearing ability in noise, and substantially reduce the difficulty those with moderate loss experience in noise. In contrast, traditional omnidirectional head-worn microphones cause a signal-to-noise deficit of about 1 dB compared to the open ear, a deficit due to the effects of head diffraction and not any particular hearing aid defect.
A little noticed advantage of directional microphones is their ability to reduce whistling caused by feedback (Knowles and Carlson, 1973, U.S. Pat. No. 3,770,911). If the ear-mold itself is well fitted, so that the vent outlet is the principal source of feedback sound, then the relationship between the vent and the microphone may sometimes be adjusted to reduce the feedback pickup by 10 or 20 dB. Similarly, the higher-performance directional microphones have a relatively low pickup to the side at high frequencies, so the feedback sound caused by faceplate vibration will see a lower microphone sensitivity than sounds coming from the front.
Despite these many advantages, the application of directional microphones has been restricted to only a small fraction of Behind-The-Ear (BTE) hearing aids, and only rarely to the much more popular In-The-Ear (ITE) hearing aids which presently comprise some 80% of all hearing aid sales.
Part of the reason for this low usage was discovered by Madafarri, who measured the diffraction about the ear and head. He found that for the same spacing between the two inlet ports of a simple first-order directional microphone, the ITE location produced only half the microphone sensitivity. Madafarri found that the diffraction of sound around the head and ear caused the effective port spacing to be reduced to about 0.7 times the physical spacing in the ITE location, while it was increased to about 1.4 times the physical spacing in the BTE location. In addition to a 2:1 sensitivity penalty for the same port spacing, the constraints of ITE hearing aid construction typically require a much smaller port spacing, further reducing sensitivity.
Another part of the reason for the low usage of directional microphones in ITE applications is the difficulty of providing the front and rear sound inlets plus a microphone cartridge in the space available. As shown in FIG. 17 of the '056 patent mentioned above, the prior art uses at least one metal inlet tube (often referred to as a nipple) welded to the side of the microphone cartridge and a coupling tube between the microphone cartridge and the faceplate of the hearing aid. The arrangement of FIG. 17 of the '056 patent wherein the microphone cartridge is also parallel with the faceplate of the hearing aide forces a spacing D as shown in that figure which may not be suitable for all ears.
A further problem is that of obtaining good directivity across frequency. Extensive experiments conducted by Madafarri as well as by applicant and his colleagues over the last 25 years have shown that in order to obtain good directivity across the audio frequencies in a head-worn directional microphone it, requires great care and a good understanding of the operation of sound in tubes (as described, for example, by Zuercher, Carlson, and Killion in their paper "Small acoustic tubes," J. Acoust. Soc. Am., V. 83, pp. 1653-1660, 1988).
A still further problem with the application of directional microphones to hearing aids is that of microphone noise. Under normal conditions, the noise of a typical non-directional hearing aid microphone cartridge is relatively unimportant to the overall performance of a hearing aid. Sound field tests show that hearing aid wearers can often detect tones within the range of 0 to 5 dB Hearing Level, i.e., within 5 dB of average young normal listeners and well within the accepted 0 to 20 dB limits of normal hearing. But when the same microphone cartridges are used to form directional microphones, a low frequency noise problem arises. The subtraction process required in first-order directional microphones results in a frequency response falling at 6 dB/octave toward low frequencies. As a result, at a frequency of 200 Hz, the sensitivity of a directional microphone may be 30 dB below the sensitivity of the same microphone cartridge operated in an omnidirectional mode.
When an equalization amplifier is used to correct the directional microphone frequency response for its low frequency drop in sensitivity, the amplifier also amplifies the low frequency noise of the microphone. In a reasonably quiet room, the amplified low frequency microphone noise may now become objectionable. Moreover, with or without equalization, the masking of the microphone noise will degrade the best aided sound field threshold at 200 Hz to approximately 35 dB HL, approaching the 40 dB HL lower limits for what is considered a moderate hearing impairment.
The equalization amplifier itself also adds to the complication of the hearing aid circuit. Thus, even in the few cases where ITE aids with directional microphones have been available, to applicant's knowledge, their frequency response has never been equalized. For this reason, Killion et al (U.S. Pat. No. 5,524,056) recommend a combination of a conventional omnidirectional microphone and a directional microphone so that the lower internal noise omnidirectional microphone may be chosen during quiet periods while the external noise rejecting directional microphone may be chosen during noisy periods.
Although directional microphones appear to be the only practical way to solve the problem of hearing in noise for the hearing-impaired individual, they have been seldom used even after nearly three decades of availability. It is the purpose of the present invention to provide an improved and fully practical directional microphone for ITE hearing aids.
Before summarizing the invention, a review of some further background information will be useful. Since the 1930s, the standard measure of performance in directional microphones has been the "directivity index" or DI, the ratio of the on-axis sensitivity of the directional microphone (sound directly in front) to that in a diffuse field (sound coming with equal probability from all directions, sometimes called random incidence sound). The majority of the sound energy at the listener's eardrum in a typical room is reflected, with the direct sound often less than 10% of the energy. In this situation, the direct path interference from a noise source located at the rear of a listener may be rejected by as much as 30 dB by a good directional microphone, but the sound reflected from the wall in front of the listener will obviously arrive from the front where the directional microphone has (intentionally) good sensitivity. If all of the reflected noise energy were to arrive from the front, the directional microphone could not help.
Fortunately, the reflections for both the desired and undesired sounds tend to be more or less random, so the energy is spread out over many arrival angles. The difference between the "random incidence" or "diffuse field" sensitivity of the microphone and its on-axis sensitivity gives a good estimate of how much help the directional microphone can give in difficult situations. An additional refinement can be made where speech intelligibility is concerned by weighing the directivity index at each frequency to the weighing function of the Articulation Index as described, for example, by Killion and Mueller on page 2 of The Hearing Journal, Vol. 43, Number 9, September 1990. Table 1 gives one set of weighing values suitable for estimating the equivalent overall improvement in signal-to-noise ratio as perceived by someone trying to understand speech in noise.
The directivity index (DI) of the two classic, first-order directional microphones, the "cosine" and "cardioid" microphones, is 4.8 dB. In the first case the microphone employs no internal acoustic time delay between the signals at the two inlets, providing a symmetrical figure 8 pattern. The cardioid employs a time delay exactly equal to the time it takes on-axis sound to travel between the two inlets. Compared to the cosine microphone, the cardioid has twice the sensitivity for sound from the front and zero sensitivity for sound from the rear. A further increase in directivity performance can be obtained by reducing the internal time delay. The hypercardioid, with minimum sensitivity for sound at 110 degrees from the front, has a DI of 6 dB. The presence of head diffraction complicates the problem of directional microphone design. For example, the directivity index for an omni BTE or ITE microphone is -1.0 to -2.0 dB at 500 and 1000 Hz.
Recognizing the problem of providing good directional microphone performance in a headworn ITE hearing aid application, applicant's set about to discover improved means and methods of such application. It is readily understood that the same solutions which make an ITE application practical can be easily applied to BTE applications as well.